What is AI governance?

Original article was published on Artificial Intelligence on Medium

An emerging field in Artificial Intelligence

Photo by Icons8 Team on Unsplash

“Governance” has, over the most recent years, become a buzz word in standard business. For the most part, governance is something that incorporates the administration. It is the manner in which rules or activities are organized, kept up, and controlled — and regularly how responsibility is relegated. Governance in artificial intelligence can be regarded as such an implementation. In general, AI governance is about AI being comprehensible, transparent, and ethical.

AI can be regarded as THE cool thing right now in the modern world. Many companies, individuals, and groups are looking into building AI integrated solutions. Yes, they can change the whole world into a better place. There are so many advanced opportunities that we can utilize from AI. Like using your face to unlock your phone without having the burden of typing a password. AI really helps to make our lives easy. However, there are some cons as well.

With the increasing number of AI integrated applications and systems being built every day, we should look into the fact of whether these applications really incorporate AI governance. In other words, are all the AI solutions explainable, transparent, and ethical?

Researchers have identified that most of the current AI solutions lack these three aspects. Most of the AI solutions tend to be biased based on their training data sets. If the developers do not consider using an unbiased sample of data for training an AI model, the final solution will obviously be biased. Moreover, AI solutions should disclose when and how AI was occupied in order to make it transparent. When the solution is transparent, there won’t be any fear of data theft, AI weaponization, or other privacy concerns. These can be regarded as highly valuable factors right now because of the increasing number of AI solutions.

Rest assured. AI governance takes care of all these factors.

A governance policy needs to control the world of AI and related bodies in order to make sure that AI solutions are really going to help mankind. Some definitions, like the Singaporean government’s, highlight that AI governance should be human-centric. Yes, but that may not be the only purpose of AI governance. It should need to be machine-centric and not optimize around humans.